Transition-Aware best-of-N sampling for Longitudinal Chest X-ray Reports

Hugging Face Daily Papers Papers

Summary

This paper introduces transition-aware best-of-N sampling, a training-free method for generating longitudinal chest X-ray reports by encoding changes between prior and current examinations using set-to-set distance metrics.

In longitudinal clinical practice, every chest X-ray is read in the context of the patients prior exam, and much of what the radiologist communicates is the change from one visit to the next. To the best of our knowledge, we present the first training-free best-of-N sampling scheme for pre-trained chest X-ray report generators that is explicitly aware of this longitudinal prior to current transition. We call it transition-aware best-of-N sampling, each report is split into sentences and embedded into an unordered set in Rd; each (prior, current) pair is reduced to a fixed-dim directional vector via a set-to-set distance designed to encode the change between the two sets; and candidates are scored by cosine distance from their candidate transition vector to a cached bank of ground-truth training transition vectors, aggregated as min or kNN. We instantiate the framework with four directional set distances (mean-shift, novelty residual, directed-Hausdorff anchor, and cost-weighted optimal transport) and evaluate on a multi-visit AP-PA cohort, running inference under three prompts on three vision-language generators. Transition-aware best-of-N outperforms random selection across the board, with the largest relative gains on the Impression section.
Original Article
View Cached Full Text

Cached at: 07/07/26, 02:44 PM

Paper page - Transition-Aware best-of-N sampling for Longitudinal Chest X-ray Reports

Source: https://huggingface.co/papers/2606.28393

Abstract

A novel training-free sampling method for chest X-ray report generation that leverages longitudinal patient history by encoding changes between prior and current examinations through set-to-set distance metrics.

In longitudinal clinical practice, every chest X-ray is read in the context of the patients prior exam, and much of what the radiologist communicates is the change from one visit to the next. To the best of our knowledge, we present the first training-freebest-of-N samplingscheme for pre-trained chest X-ray report generators that is explicitly aware of thislongitudinal priorto current transition. We call it transition-awarebest-of-N sampling, each report is split into sentences and embedded into an unordered set in Rd; each (prior, current) pair is reduced to a fixed-dimdirectional vectorvia aset-to-set distancedesigned to encode the change between the two sets; and candidates are scored bycosine distancefrom their candidatetransition vectorto a cached bank ofground-truth training transition vectors, aggregated asminorkNN. We instantiate the framework with four directional set distances (mean-shift, novelty residual, directed-Hausdorff anchor, and cost-weighted optimal transport) and evaluate on a multi-visitAP-PA cohort, running inference under three prompts on threevision-language generators. Transition-aware best-of-N outperforms random selection across the board, with the largest relative gains on the Impression section.

View arXiv pageView PDFAdd to collection

Get this paper in your agent:

hf papers read 2606\.28393

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.28393 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.28393 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.28393 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

SDR: Set-Distance Rewards for Radiology Report Generation

arXiv cs.AI

This paper proposes set-distance rewards for reinforcement learning in chest X-ray report generation, using embedding-based set-to-set distances between generated and reference reports. Post-training with these rewards via GRPO consistently outperforms supervised fine-tuning and exact-match rewards, and enables efficient test-time scaling.

Reconfigurable Radiology Labels Without Relabeling

arXiv cs.CL

This paper introduces a pipeline that converts free-text chest radiograph reports into multi-label matrices with a single structured annotation pass, enabling reconfiguration of label schemas via dictionary edits without relabeling, saving significant cost and time.